Data processing systems are indispensable in any sizeable organization today and for many organizations such as Wal-Mart they provide a major competitive information advantage. All data processing systems are primarily automated publishing systems and these provide the information that powers practical problem solving but so far they’re pretty poor publishing systems. Now a new schemaless approach to data processing based on No-SQL databases promises to prefect data processing by overcoming all the substantial publishing limitations intrinsic to the traditional data processing systems we use today.
The data processing systems we use today all too often impede problem solving because they restrict practical information flows and one way to understand this is thinking of data processing in terms of automated data clerks. In this sense every data processing program is a virtual data clerk with the powers of an automated idiot savant and blessed with a genius for accurately answering a very narrow range of highly specialized questions. In this same sense a data processing packages is a team of these highly specialized clerks but these teams seldom work well together. In a perfect world every organization would have a complete set of virtual data clerks capable of cooperating to answer any interesting practical question both accurately and adequately but it’s still far from a perfect world.
Today most organizations rely on off-the-shelf commodity data processing packages and these are typically provided by the leading enterprise software companies. Most organizations build their portfolio of data processing systems by mixing and matching products from multiple vendors and the result is always a patchwork with both duplication of package coverage and dead zones not covered by any package. Secondary manual IT systems compensate for many the dead zones while shadow IT systems built with spreadsheets and other desktop software tools compensate for other dead zones. Duplication of IT coverage requires routine maintenance and synchronization of multiple duplicate information models and these repetitive tasks involve lots of extra effort and expense.
Data processing packages are either custom built or commodity products but custom solutions are increasingly rare today because they’re both risky and ruinously expensive. Commodity packages provide only a very limited range of operational flexibility so they’re mostly a Procrustean fit for practical organization dynamics rather than a personalized fit and this poor fit all too often constrains competiveness. Many kinds of organizational processes and workflows are based on the commoditized practice models built into data processing packages but any practice that doesn’t closely conform to these cookbook models is likely to create huge IT headaches so creativity must be discouraged and conformity must be enforced. In these sense the virtual data processing clerk is capable of only very limited training and this is problematic because rapid change is constant in every competitive organization.
Data processing technology is mostly mature now and this is clearly demonstrated by many major consolidations in the data processing industry over the course of the last dozen years. Commoditization is another consequence of technological maturity and that’s why all the major data processing vendors now offer very similar product lines with very similar product features. Some data processing packages are now moving to cloud platforms but cloud product offerings are not substantially different than those available for data center hosting. Every data processing package vendor paints a rosy picture of rapid progress that will soon overcome all the frustrating limitations of traditional data processing practice but nobody much listens anymore since there’s very little actual innovation.
Data warehousing techniques are a great systematic making teams of virtual data processing clerks work together in new ways and this is particularly useful in building big picture data models of organization dynamics. Real-time data warehousing systems are making some progress today and these can overcome some of the publishing limitations of data processing systems both by supporting a broader range of question answering and by providing interesting answers to questions of somewhat greater complexity.
Now imagine a world where every organization has complete seamless data processing support for of every sort of practical activity. This is a world were data processing systems are disposable rather than durable and new fully customized data processing systems are automatically generated on demand. This is a world of schemaless information systems with far smarter information modeling and unlimited smart question answering complexity. This is the future provided by No-SQL data processing systems and it’s a future that’s just getting started today. The most important thing about this future is that it finally provides all the automated practical information flows that any organization needs and in this way overcomes all the substantial publishing limitations of the data processing systems we use today. Success in the next generation of data processing systems can only be achieved by refocusing the institutional information technology world on publishing rather than programming and this is already beginning in the best pioneering No-SQL solution projects.
The major advantages of schemaless No-SQL solutions are intrinsic simplicity and smartness. These are advantages that result in radical life-cycle costs reductions as well as information flows with far higher practical problems solving value. Schemaless information models are intrinsically sharable at Internet scale and these provide far better search success that either the data or document searching done today.
The schemaless data design cycle is remarkably simple because information models are always strictly reused rather than reinvented. The first step is preparing a sample document that contains all the kinds of practical information to be modeled and the second step is outlining the content of this document with just the same methods we master in middle school. The result of these steps is a strong schemaless information model and for many purposes this alone is sufficient to build smart codeless data processing applications. In other cases small amounts of code may be required but this coding is easily accomplished by front end web designers using familiar best practice.
How can the perfection of data processing possibly be this simple? The short answer is by eliminating all the extreme extra cost and complexity of routine information reinvention but this is heretical because it means abandoning all attempts to perfect schema-structured information modeling. The history of data modeling innovation since the 1960s can be seen as a running battle seeking to scale up schema-structured information modeling methods while struggling against the mounting costs and complexities of scale. Today this battle has finally been lost since it’s now absolutely clear that schema-structured information systems can’t possibly support mass-scale federated Internet information systems.
Schemaless information modeling is really just systematic structured document modeling with all the sharing advantages of document and all the searching advantages of data. Internet document sharing and searching is a huge Internet success story at vast scale and there’s no apparent limit to this simple scalability. The obvious way to achieve just this same success for Internet data is to exactly follow the lead of Internet document publishing and this is just where schemaless data processing is already headed. Document publishing always strictly reuses information and all schemaless data publishing can always be done in exactly the same way while preserving all the intrinsic power, precision, and portability provided by practical documents.
All traditional data processing systems have always been isolated, incommunicado information islands but in the modern Internet Age every new data processing solution should fully support freely open Internet information sharing and searching and this is a huge intrinsic advantage of schemaless data processing. Schemaless data processing solutions provide intrinsic support for plug-and-play dynamic information interoperability at Internet scale and this is a great way to easily build mass-integrated Internet data processing systems such as value chain automation systems that span whole industries. tec
Today the transition to schemaless information modeling is well underway in the Internet publishing world and this trend is driven by sheer necessity since the scale of Internet publishing increases relentlessly and only schemaless websites architectures are intrinsically saleable. Now the data processing world is just beginning to discover the big intrinsic benefits of schemaless information systems and these are the inevitable future of data processing because it’s the only future with a workable path to rapid progress and to eventual perfection.
SchemaLess.Net is the pioneering schemaless solution system provider offering lots of great schemaless information modeling software tools based on all the simple models, methods, and mindsets of mature practical information publishing practice. The SchemaLess.Net System solves all the routine publishing and editorial problems involved in building great schemaless data processing solutions that fully support scalable Internet information sharing and searching.
No comments:
Post a Comment